There is a good understanding of the genotypic basis of viral resistance to individual antiretrovirals, much of it acquired from mutational analyses of laboratory strains in cell culture. In contrast, the understanding of the genotypic basis of resistance to combination antiretroviral therapy is much weaker, particularly with respect to long-term responses. This proposal will analyze the genetic basis of resistance in these contexts by using a quantitative genetic approach. Two datasets will be analyzed. The first was collected in the Merck 035 trial of triple combination therapy, and the second will be obtained in the course of the grant from patients receiving combination therapy in Edinburgh. The data for analysis consists of 240 amino acids of the RT domain and the full 99 amino acids of the protease domain obtained at baseline of therapy. The variable sites are the analysis of amino acid sites at which the mutant amino acid is seen in at least 10% of patients. The outcome variable will be the difference between the log10 plasma viral RNA at baseline and at week 8 (or week 48) of therapy. Univariate and multiple regression will be performed on all variable amino acid sites with these outcome measures to identify the model which explains the greatest proportion of the variation in terms of baseline parameters. This analysis will generate novel information about the effect of variation in the target genes on virological responses in patients undergoing combination therapy. Early studies of RT have shown that in some cases the sites influencing long-term outcome may not previously have been identified as having any effect on response to antiretrovirals. This proposal will allow for those results to be confirmed and the approach applied to the protease gene for the first time. The applicant believes that data obtained will be of direct clinical significance for the generation of algorithms to be used in the interpretation of genotypic data on antiretroviral resistance.